| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
s. Action potentials in T cells were not significantly correlated.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The nervous system of the leech consists of a head ganglion, a chain of 21 ganglia and a tail ganglion. Each segmental ganglion innervates a specific segment of the animal (Nicholls & Baylor, 1968; Yau, 1976) and is composed of about 400 neurones, including interneurones, secretory neurones, motoneurones or sensory neurones (Macagno, 1981; Muller et al. 1981). The skin of each body segment is composed of five annuli, providing landmarks for a map of the entire skin. Coding of mechanical stimulation of the skin is primarily carried out by three kinds of mechanosensory neurones, touch (T), pressure (P) and noxious (N) cells (Nicholls & Baylor, 1968). T cells transduce light touch of the skin and each ganglion has three T cells on the left and right side. On each side of a ganglion there are two P and N cells tuned to progressively stronger stimulation of the skin (Nicholls & Baylor, 1968; Lewis & Kristan, 1998b). Mechanosensory neurones of a given ganglion innervate not only the skin of its body segment, but also that of the adjacent ganglia (Yau, 1976).
The sensitivity (Gascoigne & McVean, 1991; Carlton & McVean, 1995), receptive field characteristics (Nicholls & Baylor, 1968; Yau, 1976) and biophysical properties (Jansen & Nicholls, 1973; Van Essen, 1973; Blackshaw, 1981; Blackshaw et al. 1982; Mar & Drapeau, 1996; Pastor et al. 1996) of mechanosensory neurones have been extensively studied at a single cell level. T cells respond transiently to skin stimulation and are considered to be detectors of the mechanical environment, while P cells show tonic responses to prolonged stimulation of higher threshold and are considered responsible for triggering behaviour (Carlton & McVean, 1995). N cells are less frequently activated and have functional properties similar to the polymodal nociceptive neurones in mammals (Pastor et al. 1996). Morphological studies of the peripheral endings of these cells have established a correlation between the distribution of mechanoreceptors and the physiological properties of their receptive fields (Blackshaw, 1981; Blackshaw et al. 1982). Horseradish peroxidase injection has shown that T cell endings are the most superficial ones, being located among epithelial cells, while P and N cell endings lie deeper in the skin.
Understanding the coding of mechanical stimulation requires the characterisation of the firing pattern of all sensory neurones evoked by a given stimulus. The aim of this study was to obtain a detailed and quantitative description of sensory coding in the leech skin; in particular we want to address the following questions. (1) How many mechanosensory neurones are activated by touching the skin in a given location? (2) How reproducible and correlated is their electrical activity? (3) How does the global firing pattern change with stimulus intensity and location? (4) Is this global firing pattern dynamical? And how does it change during adaptation? In order to answer these questions it is necessary to record action potentials simultaneously from all the sensory neurones of the leech nervous system responding to a touch stimulus.
In this paper we describe experiments performed using a leech preparation that consisted of three ganglia, with a piece of body wall connected to the central ganglion. This piece of body wall consisted of five annuli, forming a single segment, and these were stimulated with appropriate devices (see Methods). Eight suction pipettes and two intracellular electrodes were used to monitor the simultaneous electrical activity of all mechanosensory neurones on the left or right side of the three ganglia. With this experimental set-up it was possible to record simultaneously the electrical activity of almost all mechanosensory neurones activated by touching the skin.
Our results indicate the basic properties of sensory coding of mechanical stimulation in the leech nervous system. When the skin is touched with a moderate tactile stimulus, i.e. exerting a force of about 20 mN, many mechanosensory neurones (T and P cells) of the three ganglia fire action potentials. If the stimulation is prolonged, many of these sensory neurones rapidly and fully adapt and only P cells in the central ganglion continue to respond. Thus sensory coding is dynamic and adaptation tunes the receptive field properties of mechanosensory neurones.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Animals and preparation
Leeches (Hirudo medicinalis) were obtained from Ricarimpex (Eysines, France) and were kept at 5°C in tap water dechlorinated by aeration for 24 h. A semi-intact preparation was used, consisting of a chain of three segmental ganglia. The central ganglion innervated the skin of either the left or right part of its body segment (see Fig. 1A). The left or right posterior and anterior roots were gently cleaned of all connective tissue, so as to allow a good en passant recording with a suction pipette with a diameter of about 80 m. The chain of ganglia and the piece of skin were pinned in a Sylgard-coated dish at room temperature (20-24°C). During dissection, the preparation was bathed in a Ringer solution with the following composition (mM): 115.0 NaCl, 1.8 CaCl2, 4.0 KCl, 12.0 glucose, 10 Tris maleate, buffered to pH 7.4 with NaOH. If a vigorous spontaneous electrical activity was recorded from all suction pipettes and if the mechanical stimulation evoked a clear electrical response in the roots and connectives, the preparation was perfused with a modified Ringer solution in which 15 mM MgCl2 replaced 15 mM NaCl, so as to block chemically mediated synaptic transmission.
Electrophysiological recording
Extracellular recordings. Eight suction pipettes were used to perform extracellular recordings as shown in Fig. 1A. Two pipettes were applied en passant onto the anterior and posterior connectives of the central ganglion. If the left (or right) part of the skin was attached to the central ganglion, two en passant suction pipettes were used to record from the left (or right) anterior and posterior roots innervating the skin. The remaining four suction pipettes were applied to the left (or right) anterior and posterior roots of adjacent ganglia.
Intracellular recordings. The electrical activity of mechanosensory neurones was also monitored by intracellular recordings (Muller et al. 1981) using sharp electrodes (input resistance around 30 M
filled with 4 M potassium acetate) and Axoclamp-2b amplifiers (Axon Instruments, Foster City, CA, USA). The extracellular voltage signals were recorded with standard analog amplifiers with a gain of 2 X 104 and a bandwidth of 200-3000 Hz. Voltage recordings were digitised at 10 kHz, stored on a personal computer and analysed with the program pCLAMP8 (Axon Instruments). Voltage signals, either intracellular or extracellular, were also stored on an 8-channel digital audio recorder (DA-88 TASCAM).
Stimulation protocol. One brief (50-80 ms) or long lasting (0.5-1 s) mechanical stimulation was delivered every 50 s to the skin by rapidly pressing nylon filaments driven by a solenoid (347-652, RS components), as described by Lewis & Kristan (1998a). Different stimulus intensities were achieved by changing the diameter and length of the filaments (Levin et al. 1978). The filament was initially placed very close to the skin so that its rapid displacement, caused by the solenoid, induced a buckling stress. The force exerted by each filament was measured by a force transducer (Honeywell FSG-15N1A) and was independent of its total buckling stress to within 15%. A stronger buckling, obtained by moving the filament closer to the skin, caused a slightly greater force. The small dependence of the force on the filament buckling was used in the experiments to determine the relative threshold of T cells in the central and adjacent ganglia.
A set of six filaments delivering stimulus intensities of about 1, 2.2, 6, 22, 50 and 100 mN was calibrated.
Neurone identification
Either at the beginning or end of the experiment all left (or right) mechanosensory neurones of the three ganglia were successively impaled with sharp intracellular electrodes. For each mechanosensory neurone, action potentials were directly evoked by passing a depolarising current pulse and by touching the skin. With this procedure it was possible to obtain a clear signature of extracellular voltage signals evoked by the action potentials in each mechanosensory neurone. These neurones have extensive branching, as shown in Fig. 1B, and extracellular voltage signals associated with action potentials can be measured by suction pipettes placed at different locations, with the amplitude of the signal depending on the size of the branch at that location. Large extracellular signals were measured from the roots branching from the ganglion in which the recording was obtained. Clear and measurable extracellular voltage signals from the connectives were recorded only from T cells; signals from P and N cells could be seen only by averaging different trials and were very small. Dorsal T and P cells (Td and Pd) produced extracellular signals only in posterior roots, while other T and P cells (Tl, Tv and Pv; v, ventral; l, lateral) had signals in both roots. Action potentials in N cells (Nm and Nl; m, medial) produced smaller voltage signals in both roots. The shapes of the extracellular signals generated by the action potentials evoked by touching the skin and by passing depolarising current into the soma were almost identical. The extracellular signals generated by an action potential from a Pv cell in the central ganglion were almost simultaneous on the anterior and posterior roots when the action potential was evoked by passing depolarising current into the cell body. However, when the action potential was evoked by touching the skin, it was first detected in the anterior root, then in the cell body after a delay of about 2 ms (see 'Intracellular recordings') and finally in the posterior root after approximately 4 ms. This difference in detection time reflects the propagation of the action potential from the endings of the anterior branch to the cell body and on to the posterior root. In some experiments when the skin near the midline was touched it was possible to observe an inversion of the detection time of signals in the two roots: sometimes the action potential wave on the anterior root preceded that on the posterior root and at other times the signals were reversed. This inversion is an indication of the variability of the locus of initiation of the action potential.
 |
View larger version [in this window] [in a new window] |
|
|
A, a photograph of three leech ganglia with a piece of skin (SKIN) connected to the central ganglion (GC). Eight suction electrodes are used to obtain en passant extracellular recordings from the anterior and posterior roots (C.G. A.R. and C.G. P.R.) and the two connectives. (P.C. and A.C.) of the central ganglion and recordings from the posterior (P.G. P.R. anD A.G. P.R.) and anterior roots (P.G. A.R. and A.G. A.R.) of the anterior (GA) and posterior (GP) ganglia. B, schematic representation of the preparation. A chain of three ganglia and the branching pattern are schematised in red, producing a typical extracellular recording. The skin is depicted by the box with graduated shading.
| ||
Action potentials elicited in T, P and N cell bodies by passing a brief depolarising current were detected by the suction pipette recording en passant from the nearest root after an average delay of 2.3 ± 0.3, 2.6 ± 0.5 and 2.8 ± 0.8 ms, respectively. This delay corresponds to the time necessary for the action potential to travel from the soma to the recording site in the root. When the skin was touched, it was possible to measure the time necessary for the action potential to travel from the roots to the cell body. This delay was on average 2.4 ± 0.4, 2.9 ± 0.6 and 3.5 ± 0.8 ms for T, P and N cells, respectively. Thus the action potentials in T, P and N cells propagate along the dendrites with similar (to within 1 ms) delays in both directions. In the case of mechanosensory neurones belonging to adjacent ganglia activated by the stimulation of their minor receptive fields, the first action potential is detected in the connective (anterior or posterior), then in the cell body (see 'Intracellular recordings') and, finally, in the roots of the anterior or posterior ganglion simultaneously.
Neurone identification was done in a semi-automatic way: different shapes of action potentials were classified as described in Pinato et al. (2000) and pairs or triplets of extracellular action potentials were detected. When different T and P cells were simultaneously active reliable neurone identification was not possible from extracellular recordings. In these cases the statistics for each neurone were recovered from intracellular recordings from visually identified mechanosensory neurones.
Receptive field characterisation
The receptive field properties of different mechanosensory neurones were analysed with brief (50 or 80 ms) or prolonged (0.5-1 s) stimulations of the skin. The piece of skin attached to the central ganglion was mapped in 12 or 20 distinct areas. This piece corresponded to the right or left hemisegment and was mapped into four regions from the dorsal to the ventral side (D, DL, VL, V) and into five regions from anterior to posterior, corresponding to the five annuli of the body segment. Each distinct area of the skin was touched with the filament at least 5 times and action potentials originating from identified mechanosensory neurones were counted and averaged in a time window of 100 ms following the stimulus onset. The receptive field was also mapped on a coarser grid with only three different regions from anterior to posterior. The receptive field profiles obtained by stimulating 12 or 20 distinct areas were the same, and therefore only 12 areas were tested in the majority of the experiments. The receptive field for a brief stimulation was constructed by counting the average number of action potentials occurring in a time window of 100 ms following the onset of the mechanical stimulation. The receptive field for a prolonged stimulation was constructed by computing the average firing frequency during the steady stimulation.
 |
View larger version [in this window] [in a new window] |
|
|
The two lowest traces are five superimposed traces from the two intracellular electrodes. Stimulation of the skin with a filament exerting a force of 50 mN for 80 ms (indicated by the top stimulus trace). Extracellular recordings as in Fig. 1A. Action potentials from identified mechanosensory neurones are indicated by different symbols (see key on the left) and colours (blue, red and green for mechanosensory neurones in the central, posterior and anterior gangla, respectively). Abbreviations defined in legend to Fig. 1.
| ||
Latency and jitter
Latency of the action potentials evoked by mechanical stimulation was computed as the delay between the onset of the mechanical stimulation and the peak of the action potential, measured either intracellularly or extracellularly. The onset of the mechanical stimulation was taken as the onset of the voltage pulse to the solenoid driving the filament displacement. The jitter of the first evoked action potential is the standard deviation of latency as defined above.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The present analysis of sensory coding and adaptation in mechanosensory neurones of the leech is based on parallel electrical recordings obtained with eight suction pipettes and two intracellular microelectrodes. To resolve action potentials in mechanosensory neurones from extracellular recordings, the preparation was perfused with a Ringer solution (Baylor & Nicholls, 1969b; Stuart, 1970). In this medium most chemical synapses were blocked and it was easy to identify action potentials originating from mechanosensory neurones.
Parallel recordings from mechanosensory neurones
Figure 2 illustrates a set of extracellular recordings obtained with eight suction pipettes when the skin was touched with a filament exerting a force of about 50 mN. Action potentials in identified mechanosensory neurones are indicated by different symbols (see key on the left of the figure) and colours (blue, red and green for mechanosensory neurones in the central, posterior and anterior ganglia, respectively). The two lowest traces are five superimposed intracellular recordings from a Pv and a Tv cell in the central ganglion. The mechanical stimulation evoked action potentials both in the central and adjacent ganglia and in T and P cells.
T cells are the first to respond to stimulation, and action potentials in these cells precede those from P and N cells by about 20 and 50 ms, respectively. These differences in latency are significantly larger than the differences in the conduction delays of T, P and N cells along the dendrites, which are similar to within 1 ms (see Methods).
The degree of correlation of the activation of T cells was analysed in experiments in which action potentials in different T cells in the central and adjacent ganglia were recorded and analysed. Figure 3A illustrates three superimposed sets of extracellular recordings from the two roots of the central, posterior and anterior ganglia, respectively, when the ventral skin was briefly touched with a filament exerting a force of about 2 mN. The skin was touched every 50 s for 47 different trials. With these experimental conditions, only action potentials from T cells were evoked and it was possible to identify a fixed number of action potentials produced by Tv cells in each trial. For instance Tv and Tl cells in the central ganglion fired five and three action potentials, respectively, in every trial and it was possible to compute the jitter of each of these action potentials. Figure 3B illustrates a single recording in which action potentials are labelled and assigned to specific T cells. The number below each identified action potential is the computed jitter in microseconds. The first three action potentials occurred at very reproducible times with a jitter of less than 100 s. All the action potentials in the Tv cell in the central ganglion had a very small jitter, always lower than 140 s. In contrast, the action potentials in the Tl cell in the central ganglion had a much larger jitter, which progressively increased with consecutive evoked action potentials. In the experiment shown in Fig. 3, the action potentials in T cells had jitters varying from just 60 s to about 1.5 ms. A similar range of jitter was observed in another seven experiments in which the occurrence of action potentials in T cells was analysed. In the great majority of the experiments analysed, a small drift of the action potential latency, of less than or about 100 s per trial, was observed. This drift was not the same for all mechanosensory neurones as one T cell could have a positive drift and another a negative one. This effect is unlikely to be caused by a drift in the stimulating apparatus, which would lead to a similar artifact in all neurones, and is probably caused by small deformations of the skin.
 |
View larger version [in this window] [in a new window] |
|
|
A, three superimposed extracellular recordings from the anterior and posterior roots of the central ganglion (two upper traces), the posterior ganglion (two middle traces) and the anterior ganglion (two lowest traces) when the skin was touched on the ventral side with a filament exerting a force equivalent to about 2 mN for 80 ms. The stimulus onset was 20 ms before the start of the trace. Forty-seven different trials were repeated every 50 s. The Tv and Tl cells in the central ganglion fired five and three action potentials, respectively, at every trial. The Tv cell in the posterior ganglion fired two action potentials and the Tv and Tl cells in the anterior ganglion fired two action potentials at every trial, of which only the first was analysed. B, one representative set of extracellular recordings showing each action potential from the T cells. Ti indicates the ith evoked action potential in the T cell. The number below each action potential is the jitter of that action potential in microseconds.
| ||
The correlation of action potentials in T cells was evaluated by computing the correlation matrix of the latencies of all identified action potentials. This matrix had positive and negative entries with absolute values usually lower than 0.5 and only occasionally entries above 0.5. While pairs of specific action potentials occasionally had a correlation around 0.7, trains of action potentials in pairs of T cells never had a correlation exceeding 0.4. These results indicate the existence of a rather small correlation between the firing of action potentials in different T cells, thus supporting the notion that the firing of action potentials in mechanosensory neurones is almost independent (see Lewis & Kristan, 1998b)
Receptive field properties with a brief mechanical stimulation
The receptive field properties of different mechanosensory neurones were analysed with brief stimulations of the skin in nine different preparations (see Methods). The piece of skin attached to the central ganglion was mapped in 12 distinct areas. This piece corresponded to the right or left hemisegment and was mapped into four regions from the dorsal to the ventral side (D, DL, VL, V) and into three regions from anterior to posterior.
Each of these areas was repeatedly stimulated and the number of action potentials occurring in a time window of 100 ms following the stimulus onset was counted and averaged over all trials. In this way a receptive field was obtained for each mechanosensory neurone from all three ganglia in the same experiment. These receptive fields, obtained with a stimulus of 100 mN, are shown in Fig. 4A-C for single T, P and N cells in central and either posterior or anterior ganglia. The receptive fields shown in Fig. 4 were obtained from a single experiment, but their spatial profiles were very similar to those obtained in all preparations analysed from nine different leeches.
 |
View larger version [in this window] [in a new window] |
|
|
T (A), P (B) and N (C) cells in the central and posterior or anterior ganglia for stimulation corresponding to 100 mN. Receptive fields were obtained by computing the average number of actio potentials in a time window of 100 ms after the onset of the stimulation. Each stimulation was repeated at least 8 times.
| ||
In all experiments aimed at mapping receptive fields, the neurones from adjacent ganglia contributed significantly to the overall evoked electrical activity. During the initial 100 ms of stimulation, the numbers of action potentials in T cells in the anterior and posterior ganglia were on average 23 ± 5 and 32 ± 7%, respectively, of those recorded from the central ganglion. For P cells these values were 16 ± 5 and 18 ± 7%. As a consequence the ratio between the number of action potentials from mechanosensory neurones in adjacent ganglia and that in the central ganglion was on average 0.55 for T cells and 0.34 for P cells.
The spatial profile of the receptive field of mechanosensory neurones from adjacent ganglia was not as smooth as that of neurones from the central ganglion and it was often possible to find a region with a very low sensitivity, surrounded by regions with appreciable sensitivities. While the receptive fields of mechanosensory neurones from the central ganglion were remarkably reproducible in different experiments, those of neurones from adjacent ganglia were variable.
Receptive fields obtained with the two stimulation strengths of about 22 and 100 mN were very similar. Therefore the spatial profile of the receptive field does not depend significantly on the intensity of the stimulus used to map it.
Threshold of action potential initiation for T cells in the central and adjacent ganglia
We have shown that the number of action potentials in T cells in adjacent ganglia is about half of that produced by T cells in the central ganglion for mechanical stimulations above 2 mN. This observation, however, was made in the presence of mechanical stimulations well above the threshold, when many action potentials were evoked. If T cells in the central ganglion and from adjacent ganglia have different thresholds, there may be mechanical stimulations for which only T cells in the central ganglion fire action potentials. In order to establish the relative threshold of T cells in the central and adjacent ganglia, the skin was touched with a filament exerting a force of about 1 mN. This force could be slightly increased or decreased by small vertical displacements of the filament, inducing different filament buckling (see Methods), so as to evoke just one action potential in a T cell in the central ganglion and/or in adjacent ganglia. The relative threshold for T cells in the central and adjacent ganglia was analysed in four preparations, in which 20 different areas were analysed.
The threshold for the activation of action potentials in T cells in the central ganglion was always smaller than that for T cells in adjacent ganglia. In 8% of cases the threshold in the posterior and anterior ganglia was the same and in 43% of cases it was lower in the anterior ganglion.
Dependence on stimulus strength with a brief mechanical stimulation
When the skin was touched with filaments exerting forces of increasing strength, different mechanosensory neurones were activated. As shown in Fig. 5A, when the skin was lightly touched with a filament exerting a force of about 1 mN on the ventral side, two action potentials in a Tv cell and one action potential in a Tl cell in the central ganglion were observed, and an action potential from a Tv cell in the anterior ganglion also appeared.
 |
View larger version [in this window] [in a new window] |
|
|
A-D, recordings from suction pipettes with stimuli exerting 1, 2.2, 6 and 22 mN, respectively. Stimulus duration was 80 ms (indicated by the top stimulus trace). Extracellular recordings as in Fig. 1A. Action potentials from identified mechanosensory neurones are labelled as in Fig. 2.
| ||
Increasing the stimulus intensity to about 2.2 mN evoked a larger number of action potentials from T cells in the central ganglion (see Fig. 5B), and with a stimulus of about 6 mN a Tv cell in the posterior ganglion was also activated (see Fig. 5C). When the stimulus intensity was increased to about 22 mN, Pv cells in the three ganglia started to fire action potentials and T cells fired more vigorously (see Fig. 5D). The dependence of the evoked activity on the stimulus intensity was analysed in nine preparations from different leeches and is shown in more detail in Fig. 6.
Figure 6A reproduces the average (7 experiments) number of action potentials, in a time window of 100 ms following the stimulus onset, as a function of stimulus intensity, for a ventral T cell (
), a ventral P cell (*) and a medial N cell (
) of the central (top) and posterior (bottom) ganglia. Panels B and C reproduce the latency and jitter, respectively, of the first evoked action potential. Data shown in Fig. 6 were averaged over seven different experiments in which the central area of the skin was touched. The threshold for activation of mechanosensory neurones in the central ganglion was very similar to that already reported (Nicholls & Baylor, 1968; Lewis & Kristan, 1998b) and was below 2 mN for T cells, about 20 mN for P cells and about 50 mN for N cells (Nicholls & Baylor, 1968). The threshold for activation of mechanosensory neurones of adjacent ganglia was slightly higher than that for neurones of the central ganglion.
The average latency of the first evoked action potential in T cells was about 50 ms for the weakest stimulation in the central ganglion and about 80 ms in the posterior ganglion, and decreased by about 30 ms when the stimulus intensity was increased from 1 to about 20 mN (see Figs 5 and 6). In one experiment, however, the latency of the first evoked action potential in a T cell in the central ganglion was always between 20 and 30 ms and did not change significantly with stimulus intensity. The first action potential evoked in P cells usually followed, by about 20 ms, the first action potential evoked in T cells and its latency did not change appreciably when the stimulus intensity was increased from 22 to 100 mN. The jitter of all mechanosensory neurones did not depend significantly on the stimulus intensity and was characteristic for each type of neurone. The jitter of the first evoked action potential was on average 1.2 ms (ranging from 0.06 to 2.6 ms) for T cells and 2.5 ms (ranging from 0.5 to 10 ms) for P cells. N cells had an average jitter of 8.5 ms, varying between 3 and 21 ms (see Fig. 6C).
 |
View larger version [in this window] [in a new window] |
|
|
A, number of evoked action potentials in a time window of 100 ms after the onset of the stimulation, as a function of stimulus intensity in the central and posterior ganglia. B, latency of the first evoked action potential as a function of stimulus intensity in the central and posterior ganglia. C, jitter (i.e. standard deviation of latency) of the first evoked action potential as a function of stimulus intensity in the central and posterior ganglia. Error bars with symbols represent standard deviation. The statistics for N cells were based on intracellular recordings and not extracellular signals, as the latter were too small to be detected during periods of intense electrical activity. Data averaged over seven different exeriments in which the central area of the skin was touched.
| ||
Responses to a prolonged stimulation
When the mechanical stimulation was prolonged up to 1 s, the recordings shown in Fig. 7 were obtained.
Panels A and C show extracellular recordings obtained when the mechanical stimulation was applied to a ventral area of the skin, as indicated in the panel insets, while panels B and D represent responses to dorsal stimulation. Recordings in response to a 22 mN stimulus are shown in panels A and B and to a 100 mN stimulus in C and D. With prolonged mechanical stimulation, T and P cells in the anterior and posterior ganglia quickly and fully adapted, while P cells in the central ganglion only slightly decreased their firing frequency. This effect was also observed in the absence of Mg2+ in the extracellular medium, thus indicating that the different adaptation of P cells in the central and adjacent ganglia is a genuine physiological property and is not caused by an alteration in the receptor sensitivity caused by Mg2+.
 |
View larger version [in this window] [in a new window] |
|
|
The two location where the stimulus was applied are indicated in the panel insets and the stimulus intensity was 22 mN in A and B and 100 mN in C and D. Stimulus duration was 1 s (indicated by the top stimulus trace). Extracellular recordings as in Fig. 1A. Action potentials from identified mechanosensory neurones are labelled as in Fig. 2.
| ||
In the presence of a stimulus corresponding to about 22 mN, P cells in the central ganglion fired at about 10 Hz, while a stronger stimulation increased the rate to about 16 Hz. T, P and N cells (see also Fig. 9D-F) responded well and reproducibly to the stimulus cessation at both stimulus intensities.
These receptive fields of mechanosensory neurones in the steady state, i.e. after development of adaptation, were very similar to those obtained with a brief stimulation (see Fig. 4), with the exception of T and P cells in adjacent ganglia, which do not fire action potentials in the steady state.
The different degree of adaptation of P cells in the central and adjacent ganglia was observed in 11 preparations, in which 12 different regions of the skin were tested. In one preparation, however, a prolonged stimulation of the skin in two specific regions did not produce a full adaptation and in the steady state a sustained discharge of 3-5 Hz was also observed in P cells in neighbouring ganglia.
Mechanisms for adaptation of T and P cells
The different degree of adaptation of P cells in the central and adjacent ganglia could be caused by a variety of biophysical mechanisms. First of all, nerve endings in the central part and in the periphery of the receptive fields may have different properties, either in the endings themselves or in the mechanical milieu. Another possibility is that action potentials do not reach the cell body of P cells in adjacent ganglia because of conduction block or reflection (Van Essen, 1973; Yau, 1976; Macagno et al. 1987; Mar & Drapeau, 1996; Baccus, 1998). In order to distinguish between these possibilities, intracellular recordings from P and N cells in the central ganglion and adjacent ganglia were performed during a prolonged stimulation.
Panels A and B of Fig. 8 reproduce intracellular recordings from the cell bodies of a Pv cell and an Nm cell, respectively, in the central ganglion during a prolonged stimulation. The Pv and Nm cells adapt to some extent and in the steady state they fire action potentials at a rate of about 16 and 6 Hz, respectively. Intracellular recordings from a Pv cell and an Nm cell of adjacent ganglia are shown in panels C and D, respectively. The Pv cell fired four action potentials and then fully adapted and stopped firing action potentials. The membrane potential of the Pv cell did not show any sign of hyperpolarisation (see Fig. 8C), typical of conduction block (Van Essen, 1973; Yau, 1976). The intracellular recording from the Nm cell shows an initial large action potential, followed by a slight membrane hyperpolarisation with some smaller depolarising deflections superimposed (see Fig. 8D), typical of conduction block (Van Essen, 1973; Macagno et al. 1987). A block occurs when the action potentials conducted by fine axons of mechanosensory neurones fail to invade large axons and the soma because of the hyperpolarisation of the cell. This behaviour was sometimes observed in N cells in adjacent ganglia, but never in P cells during a mechanical stimulation lasting 1 s, during which P cells fired at most four action potentials (see Fig. 8 and 9). When a P cell in an adjacent ganglion was repeatedly stimulated with a brief mechanical stimulation, a conduction block was observed after some seconds as previously described (Gu, 1991; Mar & Drapeau, 1996). As a consequence, central conduction block does not seem to be the major biophysical mechanism underlying full adaptation in P cells in adjacent ganglia.
Responses to two stimuli
Biophysical mechanisms underlying full adaptation of adjacent P cells were investigated in experiments using two mechanical stimuli. Stimuli 1 and 2 were applied to two different locations of the skin, L1 and L2, so that they activated different sets of nerve endings from the same P cell in an adjacent ganglion. The two locations L1 and L2 were approximately 2-3 mm apart so that the two mechanical stimulations were applied to nerve endings of the same main branch. Stimulus 1 was applied for 1 s in L1 so as to induce a full and rapid adaptation of the P cell. When adaptation was fully reached, stimulus 2 was applied in L2, so as to test the P cell sensitivity. The results of this experiment are shown in Fig. 9.
Panels A and B show extracellular recordings obtained by stimulating the two areas indicated in the insets. P cells in the central ganglion did not fully adapt, while those of adjacent ganglia fired very few action potentials and quickly and fully adapted. When the two stimuli were delivered during the same trial, the second mechanical stimulation was also able to induce action potentials in those neurones which had fully adapted to the previous stimulation, as shown in Fig. 9C. Intracellular recordings for the three stimulations used in A, B and C are shown in panels D, E and F, respectively. These recordings were obtained from Pd , Nl and Tl cells, from top to bottom, of the posterior ganglion (top and middle traces) and the central ganglion (bottom trace). In all three cells the second stimulation was able to produce action potentials. This result indicates that the full and rapid adaptation observed in mechanosensory neurones of adjacent ganglia and in T cells in the central ganglion is likely to originate from properties of nerve endings and not from conduction block or membrane properties of the cell body.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The results presented in this paper shed a new light on the coding of mechanical stimulation in the leech nervous system, by providing the first simultaneous analysis of the electrical activity evoked in mechanosensory neurones by mechanical stimulation. Several of the properties of mechanosensory neurones described here could be inferred from previous single neurone recordings (Nicholls & Baylor, 1968; Yau, 1976), but the nature of sensory coding, i.e. the relative timing, sensitivity, variability, correlation and adaptation of different mechanosensory neurones, can be firmly established only on the basis of simultaneous recordings, such as those described here.
Coding and sensitivity
When the skin was briefly touched with a thin filament exerting a weak force (above 2 mN), several T cells in the central and neighbouring ganglia fired action potentials (see Fig. 5). During the initial 100 ms, the number of action potentials in T cells in the anterior and posterior ganglia were 23 ± 5 and 32 ± 7% of those recorded from the central ganglion. T cells are able to respond to light mechanical touches exerting a force of less than or about 1 mN. Such a high sensitivity is also observed in other mechanosensitive neurones, such as those innervating hairy skin of the mouse (Koltzenburg et al. 1997). T cells in the central ganglion have a lower threshold than that of T cells in adjacent ganglia.
In agreement with previous studies (Nicholls & Baylor, 1968) when the mechanical stimulation was increased above 20 mN, P cells in the three ganglia were also stimulated. For P cells, the numbers of action potentials from the anterior and posterior ganglia were 16 ± 5 and 18 ± 7%, respectively, of those recorded from the central ganglion. The ratio between the number of action potentials from mechanosensory neurones in adjacent ganglia and that in the central ganglion was about 0.55 for T cells and 0.34 for P cells. Thus the electrical activity of adjacent ganglia is not negligible and is comparable to that of the central ganglion in terms of number of evoked action potentials. The onset of a mechanical stimulation of medium strength, i.e. about 20 mN, is signalled in the leech nervous system by T and P cells in three ganglia and by the activation of between six and twelve different mechanosensory neurones. With even stronger stimulations, corresponding to about 50 mN, N cells also respond to the stimulation (Nicholls & Baylor, 1968) and under these conditions between ten and twenty mechanosensory neurones code the mechanical stimulation.
The different sensitivity of T, P and N cells may be related to the density and architecture of their peripheral endings: indeed T cell endings are the most superficial ones (Blackshaw, 1981; Blackshaw et al. 1982) and it is not surprising that they are the most sensitive mechanosensory neurones and the first to be activated.
Timing
For any mechanical stimulation T cells are the first mechanosensory neurones to be activated. After a delay varying between 15 and 20 ms P cells also fire action potentials. Action potentials in N cells are produced with a longer delay than in T cells, varying between 40 and 80 ms. This different delay in producing action potentials is likely to originate from different mechanisms. First of all the nerve endings of T cells are the most superficial and are the first to be stimulated when the skin is touched. In addition the transduction machinery may not have the same kinetics in different mechanosensory neurones. Action potentials evoked by a mechanical stimulation occur with a delay that does not depend significantly on the stimulus strength (see Fig. 6). This behaviour is different from what is observed in invertebrate photoreceptors (Fuortes & Yeandle, 1964; Yeandle, 1985) and olfactory sensory neurones (Reisert & Matthews, 1999), where the response latency is reduced in the presence of a stronger stimulation.
Variability and correlation
Action potentials evoked in mechanosensory neurones can occur with remarkable precision and reproducibility. Action potentials evoked in T cells by a brief mechanical stimulation have a jitter varying from as low as 0.06 ms to 1 or 2 ms (see Fig. 3). Action potentials in other mechanosensory neurones have a slightly larger jitter, in the order of some milliseconds. This very low variability is not seen in invertebrate photoreceptors and olfactory sensory neurones, where highly variable bumps and events are observed (Fuortes & Yeandle, 1964; Yeandle, 1985; Menini et al. 1995), but is observed in other mechanosensory neurones in the skin of the cat and and of the monkey (Werner & Mountcastle, 1965; Vega-Bermudez & Johnson, 1999). Action potentials in T cells do not show any significant and consistent correlation in their firing pattern, despite the presence of electrical junctions among them (Baylor & Nicholls, 1969b). These results indicate that neuronal coding of mechanical stimulation has a very precise initial timing and is based on independent units.
Adaptation
In the presence of a continuous mechanical pressure, T cells in the three ganglia and P cells in neighbouring ganglia usually stop firing action potentials 200 ms after the stimulus onset (see Fig. 7) showing a rapid and full adaptation. This behaviour is reminiscent of rapidly adapting cutaneous mechanosensory neurons (Werner & Mountcastle, 1965; Burgess et al. 1983). Therefore in the presence of a weak or moderate mechanical stimulation, only P cells in the central ganglion fire action potentials. In the presence of a stronger stimulation N cells in all three ganglia also fire action potentials.
Mechanisms underlying adaptation
Different mechanisms can produce the full and complete adaptation observed in P cells in adjacent ganglia. Adaptation can occur because of the intrinsic properties of the voltage-activated conductances responsible for action potential initiation. This mechanism, however, is very unlikely to operate in our conditions, because P cells do not fully adapt when stimulated in the receptive field of their body segment (see Fig. 7). Another possible mechanism leading to adaptation of action potentials initiated in the accessory receptive fields is conduction block (Baylor & Nicholls, 1969a; Van Essen, 1973; Macagno et al. 1987). This mechanism was sometimes clearly observed in N cells (see Fig. 8D), where the first action potential initiated in the accessory field propagates into the cell body, but those following it do not. In this case the long-lasting hyperpolarisation following the somatic action potential in an N cell leads to the conduction block of action potentials arriving from distant branches (see Fig. 8D). This behaviour, however, was never observed in P cells (see Fig. 8C and 9D and E), where the first somatic action potential is not followed by a membrane hyperpolarisation with superimposed smaller depolarising waves, typical of conduction block (Yau, 1976; Gu, 1991; Mar & Drapeau, 1996; Baccus, 1998).
 |
View larger version [in this window] [in a new window] |
|
|
A and B, recordings from a Pv cell and an Nm cell, respectively, from the central ganglion. C and D, recordings from a Pv cell (anterior ganglion) and an Nm cell (posterior ganglion), respectively, during a steady stimulation lasting 1 s (indicated by the top stimulus trace). Stimulus intensity was 100 ms.
| ||
The experiments described in Fig. 9 show that a P cell in a neighbouring ganglion, which has completely and fully adapted to a prolonged mechanical stimulation of a given area of its receptive field, is able to fire action potentials in response to the stimulation of a different, but adjacent area of the receptive field. Therefore the mechanism leading to adaptation is likely to be confined to the nerve endings. The density of nerve endings in the central body segment is larger than those in the adjacent body segments (Blackshaw, 1981) and this may lead to different responses to prolonged stimulation in the centre and periphery of the receptive field. According to this view, the generator current produced during mechanotransduction at each nerve ending adapts to some extent, but the much larger density of nerve endings in the central body segment is able to induce a sustained discharge of action potentials. The lower density of nerve endings of adjacent ganglia is sufficient to produce the action potentials in the cell body at the start of the stimulus, but not after adaptation takes place at the nerve endings.
 |
View larger version [in this window] [in a new window] |
|
|
A and B, extracellular recordings showing responses to a mechanical stimulation of 1 s (A) and 0.5 s (B). C, extracellular recordings obtained when the two stimulations used in A and B were delivered simultaneously. Extracellular recordings are labelled as in Fig. 1A. D-F, intracellular recordings obtained during the same stimulation used in A-C. from a Pd cell and an Nl cell in the posterior ganglion and from a Tl cell in the central ganglion. Stimulus intensity was 100 mN. Stimulus duration indicated by top trace. Action potentials from identified mechanosensory neurones are labelled as in Fig. 2.
| ||
General properties of mechanical coding in the leech nervous system
The present analysis of the responses of mechanosensory neurones to stimulation of the skin shows some new basic properties of sensory coding in the leech skin. At the onset of the stimulation many mechanosensory neurones from at least three adjacent ganglia fire action potentials and a massive electrical signal is produced, without any clear coding of the stimulated area of the skin. After about 200 ms, this massive electrical discharge stops and a good localisation of the stimulated area is achieved by the precise coding of P cells in the central ganglion (Lewis & Kristan, 1998b). Therefore sensory coding in the leech skin is designed so as to have an initial high sensitivity but poor localisation, but within 200 ms a good degree of localisation is achieved by an appropriate gain control mechanism.
This time-dependent contraction of the receptive field properties is reminiscent of lateral inhibition observed in the eye of many invertebrate species (Coleman & Renninger, 1974; Coleman, 1975) and in the vertebrate retina (Dowling & Werblin, 1971; Werblin, 1974; Cook & McRaynolds, 1998), and of similar mechanisms observed in the cochlea (Brownell et al. 1985; Zhao & Santos-Sacchi, 1999). This dynamic control of sensitivity and localisation seems to be a basic property of sensory transduction present across species and sensory modalities.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| BACCUS D. A. (1998). Synaptic facilitation by reflected action potentials: enhancement of transmission when nerve impulses reverse direction at axon branch points. Proceedings of the National Academy of Sciences of the USA 95, 8345-8350 | |
| BAYLOR D. A. & NICHOLLS, J. G. (1969a). After-effects of nerve impulses on signalling in the central nervous system of the leech. Journal of Physiology 203, 571-589 | |
| BAYLOR D. A. & NICHOLLS, J. G. (1969b). Chemical and electrical synaptic connexions between cutaneous mechanoreceptor neurones in the central nervous system of the leech. Journal of Physiology 203, 591-609 | |
| BLACKSHAW S. E. (1981). Morphology and distribution of touch cell terminals in the skin of the leech. Journal of Physiology 320, 219-228 | |
| BLACKSHAW S. E., NICHOLLS, J. G. & PARNAS, I. (1982). Physiological responses, receptive fields and terminal arborizations of nociceptive cells in the leech. Journal of Physiology 326, 251-260 | |
| BROWNELL W. E., BADER, C. R., BERTRAND, D. & DE RIBAUPIERRE, Y. (1985). Evoked mechanical responses of isolated cochlear outer hair cells. Science 227, 194-196 | |
| BURGESS P. R., MEI, J., TUCKETT, R. P., HORCH, K. W., BALLINGER, C. M. & POULOS, D. A. (1983). The neural signal for skin indentation depth. I. Changing indentations. Journal of Neuroscience 3, 1572-1585 | |
| CARLTON T. & MCKEAN, A. (1995). The role of touch, pressure and nociceptive mechanoreceptors of the leech in unrestrained behaviour. Journal of Comparative Physiology A 177, 781-791 | |
| COLEMAN B. D. (1975). Consequences of delayed lateral inhibition in the retina of Limulus. II. Theory of spatially uniform fields, assuming the four-point property. Journal of Theoretical Biology 51, 267-291 | |
| COLEMAN B. D. & RENNINGER, G. H. (1974). Theory of delayed lateral inhibition in the compound eye of limulus. Proceedings of the National Academy of Sciences of the USA 71, 2887-2891. | |
| COOK P. B. & MCREYNOLDS, J. S. (1998). Lateral inhibition in the inner retina is important for spatial tuning of ganglion cells. Nature Neuroscience 1, 714-719 | |
| DOWLING J. E. & WERBLIN, F. S. (1971). Synaptic organization of the vertebrate retina. Vision Research 3, 1-15 | |
| FUORTES M. G. F. & YEANDLE, S. (1964). Probability of occurrence of discrete potential waves in the eye of limulus. Journal of General Physiology 47, 443-463 | |
| GASCOIGNE L. & MCVEAN, A. (1991). Neuromodulatory effects of acetylcholine and serotonin on the sensitivity of leech mechanoreceptors. Comparative Biochemistry and Physiology C 99, 369-374 | |
| GU X. (1991). Effect of conduction block at axon bifurcations on synaptic transmission to different postsynaptic neurones in the leech. Journal of Physiology 441 755-778 | |
| JANSEN J. K. S. & NICHOLLS, J. G. (1973). Conductance changes, an electrogenic pump and the hyperpolarization of the leech neurones following impulses. Journal of Physiology 229, 635-655 | |
| KOLTZENBURG M., STUCKY, C. L. & LEWIN, G. R. (1997). Receptive properties of mouse sensory neurons innervating hairy skin. Journal of Neurophysiology 78, 1841-1850 | |
| LEVIN S., PEARSALL, G. & RUDERMAN, R. J. (1978). Von Frey's method of measuring pressure sensibility in the hand: an engineering analysis of the Weinstein-Semmes pressure aesthesiometer. Journal of Hand Surgery (American) 3, 211-216 | |
| LEWIS J. E. & KRISTAN, W. B. JR (1998a). Quantitative analysis of a directed behavior in the medicinal leech: implications for organizing motor output. Journal of Neuroscience 18, 1571-1582 | |
| LEWIS J. E. & KRISTAN, W. B. JR (1998b). Representation of touch location by a population of leech sensory neurons. Journal of Neurophysiology 80, 2584-2592 | |
| MACAGNO E. R. (1981). Number and distribution of neurons in leech segmental ganglia Journal of Comparative Neurology 190, 283-302. | |
| MACAGNO E. R., MULLER, K. J. & PITMAN, R. M. (1987). Conduction block silences parts of a chemical synapse in the leech central nervous system. Journal of Physiology 387, 649-664 | |
| MAR A. & DRAPEAU, P. (1996). Modulation of conduction block in leech mechanosensory neurons. Journal of Neuroscience 16, 4335-4343 | |
| MENINI A., PICCO, C. & FIRESTEIN, S. (1995). Quantal-like current fluctuations induced by odorants in olfactory receptor cells. Nature 373, 435-437 | |
| MULLER K. J., NICHOLLS, J. G. & STENT, G. S. (1981). Neurobiology of the Leech. Cold Spring Harbor Laboratory, New York, USA. | |
| NICHOLLS J. G. & BAYLOR, D. A. (1968). Specific modalities and receptive fields of sensory neurones in CNS of the leech. Journal of Neurophysiology 31, 740-756. | |
| PASTOR J., SORIA, B. & BELMONTE, C. (1996). Properties of the nociceptive neurons of the leech segmental ganglion. Journal of Neurophysiology 75, 2268-2279 | |
| PINATO G., BATTISTON, S. & TORRE, V. (2000). Statistical independence and neural computation in the leech. Biological Cybernetics 83, 119-130 | |
| REISERT J. & MATTHEWS, H. R. (1999). Adaptation of the odour-induced response in frog olfactory receptor cells. Journal of Physiology 519, 801-813 | |
| STUART A. (1970). Physiological and morphological properties of motoneurones in the central nervous system of the leech. Journal of Physiology 209, 627-646 | |
| VAN ESSEN D. C. (1973). The contribution of membrane hyperpolarization to adaptation and conduction block in sensory neurones of the leech. Journal of Physiology 230, 509-534 | |
| VEGA-BERMUDEZ F. & JOHNSON, K. O. (1999). SA1 and RA receptive fields, response variability, and population responses mapped with a probe array. Journal of Neurophysiology 81, 2701-2710 | |
| WERBLIN F. S. (1974). Control of retinal sensitivity. II. Lateral interactions at the outer plexi form layer. Journal of General Physiology 63, 62-87 | |
| WERNER G. & MOUNTCASTLE, V. B. (1965). Neural activity in mechanoreceptive cutaneous afferents: stimulus-response relations, Weber functions and information transmission. Journal of Neurophysiology 28, 359-397 | |
| YAU K. W. (1976). Receptive fields, geometry and conduction block of sensory neurones in the central nervous system of the leech. Journal of Physiology 263, 513-538 | |
| YEANDLE S. (1985). Statistics and quantum bumps in arthropod photoreceptors. Federation Proceedings 44, 2947-2949 | |
| ZHAO H. B. & SANTOS-SACCHI, J. (1999). Auditory collusion and a coupled couple of outer hair cells. Nature 399, 359-362 |
Acknowledgements
We wish to thank Professors John Nicholls and Kenneth Muller for reading the manuscript and for useful suggestions and comments. Laura Giovanelli did the artwork. This work was supported by E.C. Project Biotech PARALLEL 960211.
Corresponding author
V. Torre: SISSA, Via Beirut 2, 34014 Trieste, Italy.
Email: torre{at}sissa.it
This article has been cited by other articles:
|
|
 |
|
  S. Andjelic and V. Torre Calcium Dynamics and Compartmentalization in Leech Neurons J Neurophysiol, December 1, 2005; 94(6): 4430 - 4440. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Baca, E. E. Thomson, and W. B. Kristan Jr. Location and Intensity Discrimination in the Leech Local Bend Response Quantified Using Optic Flow and Principal Components Analysis J Neurophysiol, June 1, 2005; 93(6): 3560 - 3572. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Garcia-Perez, D. Zoccolan, G. Pinato, and V. Torre Dynamics and Reproducibility of a Moderately Complex Sensory-Motor Response in the Medicinal Leech J Neurophysiol, September 1, 2004; 92(3): 1783 - 1795. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Zoccolan, G. Pinato, and V. Torre Highly Variable Spike Trains Underlie Reproducible Sensorimotor Responses in the Medicinal Leech J. Neurosci., December 15, 2002; 22(24): 10790 - 10800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Zoccolan and V. Torre Using Optical Flow to Characterize Sensory-Motor Interactions in a Segment of the Medicinal Leech J. Neurosci., March 15, 2002; 22(6): 2283 - 2298. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
